
#####################################################################################################


#####################################################################################################


import os
import cv2
import numpy as np
import tensorflow as tf
from keras.models import model_from_json
from keras.preprocessing import image
import matplotlib.pyplot as plt
import matplotlib.image as mpimg


# (0=Angry, 1=Disgust, 2=Fear, 3=Happy, 4=Sad, 5=Surprise, 6=Neutral)


model = model_from_json(open("model.json", "r").read())
model.load_weights('model.h5') 
face_haar_cascade = cv2.CascadeClassifier(cv2.data.haarcascades + 'haarcascade_frontalface_default.xml')


cap=cv2.VideoCapture(0)  #select the default video capture

#If the camera was not opened sucessfully
if not cap.isOpened():  
    print("Cannot open camera")
    exit()
    
#Continously read the frames 
while True:
    #read frame by frame and get return whether there is a stream or not
    ret, frame=cap.read()
    
    #If no frames recieved, then break from the loop
    if not ret:
        print("Can't receive frame (stream end?). Exiting ...")
        break
        
    #Change the frame to greyscale  
    gray_image= cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    #We pass the image, scaleFactor and minneighbour
    faces_detected = face_haar_cascade.detectMultiScale(gray_image,1.32,5)
    
    #Draw Triangles around the faces detected
    for (x,y,w,h) in faces_detected:
        cv2.rectangle(frame,(x,y), (x+w,y+h), (255,0,0), thickness=7)
        roi_gray=gray_image[y:y+w,x:x+h]
        roi_gray=cv2.resize(roi_gray,(48,48))
        
        #Processes the image and adjust it to pass it to the model
        image_pixels = tf.keras.preprocessing.image.img_to_array(roi_gray)
        plt.imshow(image_pixels)
        plt.show()
        image_pixels = np.expand_dims(image_pixels, axis = 0)
        image_pixels /= 255

        #Get the prediction of the model
        predictions = model.predict(image_pixels)
        print(predictions)
        max_index = np.argmax(predictions[0])
        emotion_detection = ('angry', 'disgust', 'fear', 'happy', 'sad', 'surprise', 'neutral')
        emotion_prediction = emotion_detection[max_index]
        
        print('Prediction: Right Now you are: ' + emotion_prediction)

        angry_rate = np.take(predictions[0],0)*100
        disgusted_rate = np.take(predictions[0],1)*100
        fear_rate = np.take(predictions[0],2)*100
        happy_rate = np.take(predictions[0],3)*100
        sad_rate = np.take(predictions[0],4)*100
        surprise_rate = np.take(predictions[0],5)*100
        neutral_rate = np.take(predictions[0],6)*100
        
        print('---- Angry: '+str(angry_rate)+'%')
        print('---- Disgust: '+str(disgusted_rate)+'%')
        print('---- Fear: '+str(fear_rate)+'%')
        print('---- Happy: '+str(happy_rate)+'%')
        print('---- Sad: '+str(sad_rate)+'%')
        print('---- Surprise: '+str(surprise_rate)+'%')
        print('---- Neutral: '+str(neutral_rate)+'%')
        
        #Write on the frame the emotion detected
        cv2.putText(frame,emotion_prediction,(int(x), int(y)),cv2.FONT_HERSHEY_SIMPLEX,2,(0,255,0),3)
    
    
    resize_image = cv2.resize(frame, (1000, 700))
    cv2.imshow('Emotion',resize_image)
    if cv2.waitKey(1) & 0xFF == ord('q'):
            break


cap.release()
cv2.destroyAllWindows

1/1 [==============================] - 0s 220ms/step
[[0.07805975 0.00128663 0.26617336 0.05065374 0.5310841  0.0063972
  0.06634526]]
Prediction: Right Now you are: sad
---- Angry: 7.805974781513214%
---- Disgust: 0.12866259785369039%
---- Fear: 26.61733627319336%
---- Happy: 5.065374076366425%
---- Sad: 53.10841202735901%
---- Surprise: 0.6397201679646969%
---- Neutral: 6.634525954723358%

1/1 [==============================] - 0s 85ms/step
[[0.1941406  0.00653147 0.49993575 0.04797022 0.17068629 0.0197426
  0.06099311]]
Prediction: Right Now you are: fear
---- Angry: 19.414059817790985%
---- Disgust: 0.6531465798616409%
---- Fear: 49.99357461929321%
---- Happy: 4.797021672129631%
---- Sad: 17.06862896680832%
---- Surprise: 1.974259503185749%
---- Neutral: 6.099310517311096%

1/1 [==============================] - 0s 49ms/step
[[0.20279175 0.00627842 0.45076197 0.043662   0.18802465 0.01842347
  0.09005772]]
Prediction: Right Now you are: fear
---- Angry: 20.27917504310608%
---- Disgust: 0.627841567620635%
---- Fear: 45.076197385787964%
---- Happy: 4.366200417280197%
---- Sad: 18.802465498447418%
---- Surprise: 1.8423466011881828%
---- Neutral: 9.005771577358246%

1/1 [==============================] - 0s 61ms/step
[[0.41078228 0.01023022 0.34988296 0.04073784 0.11152364 0.01705601
  0.059787  ]]
Prediction: Right Now you are: angry
---- Angry: 41.07822775840759%
---- Disgust: 1.0230219922959805%
---- Fear: 34.988296031951904%
---- Happy: 4.073784127831459%
---- Sad: 11.152363568544388%
---- Surprise: 1.705600880086422%
---- Neutral: 5.978699773550034%

1/1 [==============================] - 0s 61ms/step
[[0.18280098 0.00451663 0.68596274 0.01422935 0.07633052 0.01122672
  0.02493305]]
Prediction: Right Now you are: fear
---- Angry: 18.28009784221649%
---- Disgust: 0.4516633227467537%
---- Fear: 68.59627366065979%
---- Happy: 1.4229352585971355%
---- Sad: 7.63305202126503%
---- Surprise: 1.1226717382669449%
---- Neutral: 2.4933049455285072%

1/1 [==============================] - 0s 50ms/step
[[0.08268981 0.00133126 0.530022   0.00866766 0.33583367 0.00342652
  0.03802902]]
Prediction: Right Now you are: fear
---- Angry: 8.26898142695427%
---- Disgust: 0.13312576338648796%
---- Fear: 53.00220251083374%
---- Happy: 0.8667662739753723%
---- Sad: 33.58336687088013%
---- Surprise: 0.3426517127081752%
---- Neutral: 3.80290225148201%

1/1 [==============================] - 0s 34ms/step
[[0.12236146 0.00222143 0.3955433  0.01595321 0.3794627  0.00522211
  0.07923581]]
Prediction: Right Now you are: fear
---- Angry: 12.2361458837986%
---- Disgust: 0.2221428556367755%
---- Fear: 39.554330706596375%
---- Happy: 1.595320738852024%
---- Sad: 37.946268916130066%
---- Surprise: 0.5222107749432325%
---- Neutral: 7.923580706119537%

1/1 [==============================] - 0s 49ms/step
[[6.4423174e-02 4.6446754e-04 1.8297356e-01 1.7747741e-02 6.9216454e-01
  1.6457831e-03 4.0580679e-02]]
Prediction: Right Now you are: sad
---- Angry: 6.442317366600037%
---- Disgust: 0.04644675354938954%
---- Fear: 18.297356367111206%
---- Happy: 1.774774119257927%
---- Sad: 69.21645402908325%
---- Surprise: 0.16457830788567662%
---- Neutral: 4.058067873120308%

1/1 [==============================] - 0s 43ms/step
[[9.5212013e-03 2.0010541e-05 3.6293551e-01 8.2081527e-04 6.1830509e-01
  1.3137201e-04 8.2660001e-03]]
Prediction: Right Now you are: sad
---- Angry: 0.9521201252937317%
---- Disgust: 0.002001054053835105%
---- Fear: 36.293551325798035%
---- Happy: 0.08208152721635997%
---- Sad: 61.83050870895386%
---- Surprise: 0.013137201312929392%
---- Neutral: 0.826600007712841%

1/1 [==============================] - 0s 38ms/step
[[2.1917725e-03 2.2166692e-07 3.3129875e-02 1.9245053e-04 9.6298671e-01
  2.4559458e-06 1.4965945e-03]]
Prediction: Right Now you are: sad
---- Angry: 0.21917724516242743%
---- Disgust: 2.216669230392654e-05%
---- Fear: 3.3129874616861343%
---- Happy: 0.019245053408667445%
---- Sad: 96.2986707687378%
---- Surprise: 0.0002455945832480211%
---- Neutral: 0.14965945156291127%

PFA: Facial expression recognition

Goals:

Additional functions we can add:

Real Time Detection¶